CWE-1426 Base Incomplet

Improper Validation of Generative AI Output

This vulnerability occurs when an application uses a generative AI model (like an LLM) but fails to properly check the AI's output before using it. Without this validation, the AI's responses might…

Définition

What is CWE-1426?

This vulnerability occurs when an application uses a generative AI model (like an LLM) but fails to properly check the AI's output before using it. Without this validation, the AI's responses might contain security flaws, harmful content, or data leaks that violate the application's intended policies.
Generative AI models are powerful but unpredictable. They can be tricked into producing malicious code, biased decisions, offensive content, or sensitive training data. If your application blindly trusts and acts on these outputs, it can lead to injection attacks, compliance violations, or data breaches. You must implement robust validation checks—like content filtering, code sanitization, and policy enforcement—on every AI response before it's processed further. Continuously monitoring for these validation failures across all your AI-integrated services is a complex challenge. An ASPM platform like Plexicus can automatically detect these flaws in your runtime environment, while its AI-powered remediation provides specific fixes to harden your validation logic, ensuring your AI features remain secure and reliable.
Impact réel

Real-world CVEs caused by CWE-1426

  • CVE-2024-3402

    chain: GUI for ChatGPT API performs input validation but does not properly "sanitize" or validate model output data (CWE-1426), leading to XSS (CWE-79).

Comment les attaquants l'exploitent

Parcours de l'attaquant étape par étape

  1. 1

    Identifier un chemin de code qui traite des entrées non fiables sans validation.

  2. 2

    Élaborer une charge utile qui exploite le comportement non sécurisé — injection, traversal, débordement ou abus de logique.

  3. 3

    Délivrer la charge utile via une requête normale et observer la réaction de l'application.

  4. 4

    Itérer jusqu'à ce que la réponse divulgue des données, exécute le code de l'attaquant ou élève les privilèges.

Exemple de code vulnérable

Vulnerable pseudo

MITRE n'a pas publié d'exemple de code pour ce CWE. Le motif ci-dessous est illustratif — voir Ressources pour les références canoniques.

Vulnérable pseudo 
// Example pattern — see MITRE for the canonical references.
function handleRequest(input) {
  // Untrusted input flows directly into the sensitive sink.
  return executeUnsafe(input);
}
Exemple de code sécurisé

Secure pseudo

Sécurisé pseudo 
// Validate, sanitize, or use a safe API before reaching the sink.
function handleRequest(input) {
  const safe = validateAndEscape(input);
  return executeWithGuards(safe);
}
What changed: the unsafe sink is replaced (or the input is validated/escaped) so the same payload no longer triggers the weakness.
Liste de contrôle de prévention

How to prevent CWE-1426

  • Architecture and Design Since the output from a generative AI component (such as an LLM) cannot be trusted, ensure that it operates in an untrusted or non-privileged space.
  • Operation Use "semantic comparators," which are mechanisms that provide semantic comparison to identify objects that might appear different but are semantically similar.
  • Operation Use components that operate externally to the system to monitor the output and act as a moderator. These components are called different terms, such as supervisors or guardrails.
  • Build and Compilation During model training, use an appropriate variety of good and bad examples to guide preferred outputs.
Signaux de détection

How to detect CWE-1426

Dynamic Analysis with Manual Results Interpretation

Use known techniques for prompt injection and other attacks, and adjust the attacks to be more specific to the model or system.

Dynamic Analysis with Automated Results Interpretation

Use known techniques for prompt injection and other attacks, and adjust the attacks to be more specific to the model or system.

Architecture or Design Review

Review of the product design can be effective, but it works best in conjunction with dynamic analysis.

Correction automatique Plexicus

Plexicus détecte automatiquement CWE-1426 et ouvre une PR de correction en moins de 60 secondes.

Codex Remedium analyse chaque commit, identifie cette faiblesse précise et livre une pull request prête à être relue avec le correctif. Pas de tickets. Pas de transferts.

Obtenir une démo Essayer Plexicus gratuitement
Questions fréquentes

Frequently asked questions

Qu'est-ce que CWE-1426 ?

This vulnerability occurs when an application uses a generative AI model (like an LLM) but fails to properly check the AI's output before using it. Without this validation, the AI's responses might contain security flaws, harmful content, or data leaks that violate the application's intended policies.

Quelle est la gravité de CWE-1426 ?

MITRE n'a pas publié de note de probabilité d'exploitation pour cette faiblesse. Traitez-la comme un impact moyen jusqu'à ce que votre modèle de menace prouve le contraire.

Quels langages ou plateformes sont affectés par CWE-1426 ?

MITRE lists the following affected platforms: Not Architecture-Specific, AI/ML, Not Technology-Specific.

Comment puis-je prévenir CWE-1426 ?

Since the output from a generative AI component (such as an LLM) cannot be trusted, ensure that it operates in an untrusted or non-privileged space. Use "semantic comparators," which are mechanisms that provide semantic comparison to identify objects that might appear different but are semantically similar.

Comment Plexicus détecte et corrige CWE-1426 ?

Le moteur SAST de Plexicus reconnaît la signature de flux de données de CWE-1426 à chaque commit. Lorsqu'une correspondance est trouvée, notre agent Codex Remedium ouvre une PR de correction avec le code corrigé, les tests et un résumé d'une ligne pour le relecteur.

Où puis-je en savoir plus sur CWE-1426 ?

MITRE publie la définition canonique à https://cwe.mitre.org/data/definitions/1426.html. Vous pouvez également consulter la documentation OWASP et NIST pour des conseils adjacents.

Faiblesses associées

Weaknesses related to CWE-1426

CWE-707 Parent

Improper Neutralization

This vulnerability occurs when an application fails to properly validate or sanitize structured data before it's received from an external…

CWE-116 Frère

Improper Encoding or Escaping of Output

This vulnerability occurs when an application builds a structured message—like a query, command, or request—for another component but…

CWE-138 Frère

Improper Neutralization of Special Elements

This vulnerability occurs when an application accepts external input but fails to properly sanitize special characters or syntax that have…

CWE-170 Frère

Improper Null Termination

This weakness occurs when software fails to properly end a string or array with the required null character or equivalent terminator.

CWE-172 Frère

Encoding Error

This vulnerability occurs when software incorrectly transforms data between different formats, leading to corrupted or misinterpreted…

CWE-182 Frère

Collapse of Data into Unsafe Value

This vulnerability occurs when an application's data filtering or transformation process incorrectly merges or simplifies information,…

CWE-20 Frère

Improper Input Validation

This vulnerability occurs when an application accepts data from an external source but fails to properly verify that the data is safe and…

CWE-228 Frère

Improper Handling of Syntactically Invalid Structure

This vulnerability occurs when software fails to properly reject or process input that doesn't follow the expected format or structure,…

CWE-240 Frère

Improper Handling of Inconsistent Structural Elements

This vulnerability occurs when a system fails to properly manage situations where related data structures or elements should match but are…

Ressources

Further reading

Prêt quand vous l'êtes

Arrêtez de payer par développeur.
Commencez à fermer la boucle.

Plexicus est l'ASPM natif IA qui scanne, filtre, corrige, penteste et explique — de façon autonome. Développeurs illimités, dépôts illimités, actions IA à usage équitable. Vrai niveau gratuit, €269/mo annuel quand vous êtes prêt.

Commencer gratuitement Réserver une démo